A Comprehensive Study of Activity Recognition Using Accelerometers

This paper serves as a survey and empirical evaluation of the state-of-the-art in activity recognition methods using accelerometers. The paper is particularly focused on long-term activity recognition in real-world settings. In these environments, data collection is not a trivial matter; thus, there...

Full description

Bibliographic Details
Main Authors: Niall Twomey, Tom Diethe, Xenofon Fafoutis, Atis Elsts, Ryan McConville, Peter Flach, Ian Craddock
Format: Article
Language:English
Published: MDPI AG 2018-05-01
Series:Informatics
Subjects:
Online Access:http://www.mdpi.com/2227-9709/5/2/27
id doaj-87177d4613a14330a5ce4b4461a56fee
record_format Article
spelling doaj-87177d4613a14330a5ce4b4461a56fee2020-11-24T22:30:31ZengMDPI AGInformatics2227-97092018-05-01522710.3390/informatics5020027informatics5020027A Comprehensive Study of Activity Recognition Using AccelerometersNiall Twomey0Tom Diethe1Xenofon Fafoutis2Atis Elsts3Ryan McConville4Peter Flach5Ian Craddock6School of Computer Science, Electrical and Electronic Engineering, and Engineering Mathematics, University of Bristol, Bristol BS8 1UB, UKSchool of Computer Science, Electrical and Electronic Engineering, and Engineering Mathematics, University of Bristol, Bristol BS8 1UB, UKSchool of Computer Science, Electrical and Electronic Engineering, and Engineering Mathematics, University of Bristol, Bristol BS8 1UB, UKSchool of Computer Science, Electrical and Electronic Engineering, and Engineering Mathematics, University of Bristol, Bristol BS8 1UB, UKSchool of Computer Science, Electrical and Electronic Engineering, and Engineering Mathematics, University of Bristol, Bristol BS8 1UB, UKSchool of Computer Science, Electrical and Electronic Engineering, and Engineering Mathematics, University of Bristol, Bristol BS8 1UB, UKSchool of Computer Science, Electrical and Electronic Engineering, and Engineering Mathematics, University of Bristol, Bristol BS8 1UB, UKThis paper serves as a survey and empirical evaluation of the state-of-the-art in activity recognition methods using accelerometers. The paper is particularly focused on long-term activity recognition in real-world settings. In these environments, data collection is not a trivial matter; thus, there are performance trade-offs between prediction accuracy, which is not the sole system objective, and keeping the maintenance overhead at minimum levels. We examine research that has focused on the selection of activities, the features that are extracted from the accelerometer data, the segmentation of the time-series data, the locations of accelerometers, the selection and configuration trade-offs, the test/retest reliability, and the generalisation performance. Furthermore, we study these questions from an experimental platform and show, somewhat surprisingly, that many disparate experimental configurations yield comparable predictive performance on testing data. Our understanding of these results is that the experimental setup directly and indirectly defines a pathway for context to be delivered to the classifier, and that, in some settings, certain configurations are more optimal than alternatives. We conclude by identifying how the main results of this work can be used in practice, specifically in experimental configurations in challenging experimental conditions.http://www.mdpi.com/2227-9709/5/2/27activities of daily livingactivity recognitionaccelerometersmachine learningsensors
collection DOAJ
language English
format Article
sources DOAJ
author Niall Twomey
Tom Diethe
Xenofon Fafoutis
Atis Elsts
Ryan McConville
Peter Flach
Ian Craddock
spellingShingle Niall Twomey
Tom Diethe
Xenofon Fafoutis
Atis Elsts
Ryan McConville
Peter Flach
Ian Craddock
A Comprehensive Study of Activity Recognition Using Accelerometers
Informatics
activities of daily living
activity recognition
accelerometers
machine learning
sensors
author_facet Niall Twomey
Tom Diethe
Xenofon Fafoutis
Atis Elsts
Ryan McConville
Peter Flach
Ian Craddock
author_sort Niall Twomey
title A Comprehensive Study of Activity Recognition Using Accelerometers
title_short A Comprehensive Study of Activity Recognition Using Accelerometers
title_full A Comprehensive Study of Activity Recognition Using Accelerometers
title_fullStr A Comprehensive Study of Activity Recognition Using Accelerometers
title_full_unstemmed A Comprehensive Study of Activity Recognition Using Accelerometers
title_sort comprehensive study of activity recognition using accelerometers
publisher MDPI AG
series Informatics
issn 2227-9709
publishDate 2018-05-01
description This paper serves as a survey and empirical evaluation of the state-of-the-art in activity recognition methods using accelerometers. The paper is particularly focused on long-term activity recognition in real-world settings. In these environments, data collection is not a trivial matter; thus, there are performance trade-offs between prediction accuracy, which is not the sole system objective, and keeping the maintenance overhead at minimum levels. We examine research that has focused on the selection of activities, the features that are extracted from the accelerometer data, the segmentation of the time-series data, the locations of accelerometers, the selection and configuration trade-offs, the test/retest reliability, and the generalisation performance. Furthermore, we study these questions from an experimental platform and show, somewhat surprisingly, that many disparate experimental configurations yield comparable predictive performance on testing data. Our understanding of these results is that the experimental setup directly and indirectly defines a pathway for context to be delivered to the classifier, and that, in some settings, certain configurations are more optimal than alternatives. We conclude by identifying how the main results of this work can be used in practice, specifically in experimental configurations in challenging experimental conditions.
topic activities of daily living
activity recognition
accelerometers
machine learning
sensors
url http://www.mdpi.com/2227-9709/5/2/27
work_keys_str_mv AT nialltwomey acomprehensivestudyofactivityrecognitionusingaccelerometers
AT tomdiethe acomprehensivestudyofactivityrecognitionusingaccelerometers
AT xenofonfafoutis acomprehensivestudyofactivityrecognitionusingaccelerometers
AT atiselsts acomprehensivestudyofactivityrecognitionusingaccelerometers
AT ryanmcconville acomprehensivestudyofactivityrecognitionusingaccelerometers
AT peterflach acomprehensivestudyofactivityrecognitionusingaccelerometers
AT iancraddock acomprehensivestudyofactivityrecognitionusingaccelerometers
AT nialltwomey comprehensivestudyofactivityrecognitionusingaccelerometers
AT tomdiethe comprehensivestudyofactivityrecognitionusingaccelerometers
AT xenofonfafoutis comprehensivestudyofactivityrecognitionusingaccelerometers
AT atiselsts comprehensivestudyofactivityrecognitionusingaccelerometers
AT ryanmcconville comprehensivestudyofactivityrecognitionusingaccelerometers
AT peterflach comprehensivestudyofactivityrecognitionusingaccelerometers
AT iancraddock comprehensivestudyofactivityrecognitionusingaccelerometers
_version_ 1725740563913768960